Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Silberman, Bentin and Miikkulainen 1
Semantic Boost on Episodic Associations: An Empirically-based
Computational Model
Yaron Silberman Department of Psychology and Interdisciplinary Center for Neural Computation,
The Hebrew University of Jerusalem, Giv'at Ram, Jerusalem 91904 Israel
Shlomo Bentin Department of Psychology and Interdisciplinary Center for Neural Computation,
The Hebrew University of Jerusalem, Giv'at Ram, Jerusalem 91904 Israel
Risto Miikkulainen Department of Computer Science, The University of Texas at Austin
Austin, TX 78712-1188 USA
Address for editorial correspondence:
Risto Miikkulainen Department of Computer Sciences The University of Texas at Austin 1 University Station C0500 Austin, TX 78712 E:Mail: [email protected] Key Words: associations, episodic memory, semantic memory, neural networks, semantic
maps
mailto:[email protected]
Silberman, Bentin and Miikkulainen 2
Abstract
Humans associate words remarkably easily and efficiently. Frequently the
associations are formed incidentally and such associations provide a window into the
cognitive structure of language networks and memory. How associations between
words are formed, however, is not well understood. In a series of incidental
associative learning experiments human participants associated semantically related
but unassociated words easier than unrelated words. Furthermore, this advantage
increased linearly as a function of the number of repeated co-occurrences of the word-
pairs, up to a ceiling. In order to understand the mechanism of the semantic influence
on word association processes a computational model, SEMANT, was developed. In
this model, episodic associations are implemented by lateral connections that are
formed between nodes in an already existing self-organized semantic map. These
lateral connections are strengthened by each concomitant activation of two nodes in
direct proportion with the amount of overlapping activity. Simulating associative
learning with SEMANT replicated the empirical findings and led to two testable
predictions. First, it should be easier to associate a word with few semantic neighbors
to a word with many semantic neighbors than vice versa. Second, diffuse semantic
activation, such as that characteristic to schizophrenic thought disorders, should
reduce the advantage of associating semantically related pairs over unrelated pairs.
The SEMANT model, therefore, suggests how the interaction of episodic and
semantic memory can be understood in terms of specific neural computations.
Silberman, Bentin and Miikkulainen 3
1. Introduction
Word associations are basic building blocks of human cognition in general and
human learning in particular. Understanding how such associations are formed can
give us a fundamental insight into the cognitive structures underlying learning and
memory.
A viable association between words is demonstrated phenomenologically
when the presentation of one word brings the second word to the perceiver's
awareness. By definition, associations are based primarily on episodic learning. That
is, words are associated if they co-occur in time and space, and the strength of an
association is determined by the cumulative effect of such episodes. An important
aspect of this definition is that associative learning is not necessarily intentional.
Indeed, associations are often established incidentally, that is, without intention and
without allocating attention to the learning process.
Connections between words can also be based on a semantic relationship. For
example, words can share common semantic features (e.g. when they belong to the
same semantic category such as goat - lion), can maintain a part-whole relationship
(e.g. wheel - car), a functional relationship (e.g. broom - floor) and more (Cruse,
1986). Note that if words co-occur frequently they may become associated whether
they are semantically related or not, but semantically related words are not always
associated (Prior & Geffet, 2003; Neely, 1991). Indeed, episodic associations and
semantic relationships are probably based on different properties. Yet, many
semantically related word pairs are also episodically associated, which might suggest
that the two types of relations, indeed, interact.
Silberman, Bentin and Miikkulainen 4
This paper focuses on understanding this interaction in detail. Performance
experiments with human participants were first performed to characterize the
interaction between the categorical neighborhood organization of the semantic system
and the episodic dynamics of associative learning. In these experiments the
participants' ability to form associations between semantically related and unrelated
words was compared. Semantic relations were found to facilitate forming new
associations in a specific manner: They provide only a negligible initial advantage,
but instead facilitate the process by which the episodic associations are strengthened
over multiple repetitions, up to a ceiling.
These performance patterns were then used as a starting point for developing a
computational theory of word associations, expressed in the SEMANT computational
model. In SEMANT, semantic similarity is based on distributed representations (cf.
HAL) and episodic associations are based on spreading activation, both represented
together in a laterally connected semantic map. The model was validated in simulated
associative learning experiments, and then used to draw predictions for future human
experiments on asymmetry of associations and potential causes of mediated
associations in schizophrenia. In this manner, SEMANT expresses a computational
theory of semantic and episodic effects in word associations, and serves to drive
future research.
2. Background
There is considerable evidence that semantic and episodic memory systems
interact. Most of such studies (e.g. McKoon & Ratcliff, 1986; Neely & Durgunoglu,
1985; Shelton & Martin, 1992) were based on the semantic priming paradigm and
focused on examining existing structures. In particular, several studies showed
considerably larger semantic priming effects on performance if the prime and the
Silberman, Bentin and Miikkulainen 5
target are also episodically associated (Chiarello et al. 1990; McRae & Boisvert,
1998); to some extent, these results generalize into newly formed associations as well
(Dagenbach, Horst, & Car, 1990; Schrijnemakers & Raaijmakers, 1997).
A converse influence of semantic relationship on associative learning has also
been documented. For example, early studies reported more accurate cued recall for
strongly related than for weakly related word pairs using the paired associate
paradigm (e.g., McGuire, 1961; Underwood & Schultz, 1960; Wicklund, Palermo, &
Jenkins, 1964). Supporting these earlier findings, Epstein, Phillips & Johnson (1975)
suggested that semantically related words were more likely to be associated during an
incidental learning procedure than unrelated pairs, even if the orientation task did not
involve semantic processing. More recently, Guttentag (1995) reported similar
findings in children. Moreover, using sentential context rather than relatedness
between words, Prior & Bentin (2003) found that associations between unrelated
words are formed more easily when they are embedded in a sentence than when they
are presented as isolated pairs.
Although such studies suggest that semantic relationships between words
affect how associations are formed between them, the data needs additional
corroboration, primarily because the strength of pre-experimental associative links
between semantically related words has not been sufficiently controlled. Hence, the
reported advantage for associating semantically related over unrelated words could
have been induced by associative connections that existed before the experiment
rather than by an interaction between the two types of connections during associative
learning. Therefore, human experiments investigating the semantic influence on the
Silberman, Bentin and Miikkulainen 6
formation of new associations between words were performed as a starting point in
the current study, as will be described in the next section.
Given that empirical data exists, computational modeling is an ideal tool to
make sense of it and to formulate specific theories about the process. The modeling
can be done at several different levels of abstraction, depending on the level of
explanation desired. At the highest level, the semantic and episodic interactions can
be formalized in Bower's one-element model (Bower, 1961). For example, the
semantic relationship advantage could be modeled so that in each trial there's a fixed
chance that an association is learned between semantically related word pairs, and a
smaller fixed chance that it is learned between semantically unrelated pairs. Such a
model accounts for the interaction, but it does not explain how these learning patterns
emerge from memory structures, and it is therefore limited in its predictive power.
At a more detailed level, a popular approach to explaining the organization of
semantic memory is representing concepts by distinguishable patterns of activity over
a large number of nodes (Farah & McClelland, 1991; Hinton, 1990; Hinton &
Shallice, 1991; Masson, 1995; Moss et al., 1994; Plaut, 1995; Plaut & Shallice, 1993).
Each node participating in a representation accounts for a specific semantic micro-
feature, and semantic similarity is expressed as an overlap in activity patterns over the
set of micro-features. Activity propagates through recurrent connections until the
network settles to a stable state (an attractor), representing the semantically related
items. Such a model matches much of the same data but allows more detailed
explanations and predictions to be drawn. For example, Plaut's (1995) model suggests
a distinction between the way by which semantic and episodic relations are formed:
Whereas semantic relations are encoded in the micro-features, episodic relations
Silberman, Bentin and Miikkulainen 7
between word pairs are based on co-occurrence information throughout the learning
history. On the other hand, the attractor models are rather difficult to interpret, and it
is difficult to separate the different factors contributing to their behavior. In particular,
Plaut's (1995) model does not specify what mechanisms underlie the process of
forming episodic associations and how the strength of existing semantic connections
among other factors affects this process.
The goal of computational modeling, however, is not just to model the data,
but to be able to draw predictions that can be verified in future empirical experiments.
The SEMANT model is a principled model specifically designed to suggest and test
such a mechanism. SEMANT is not an attractor network, but is instead based on three
well-known principles that make its processing and learning transparent: map-like
memory organization, spreading activation, and Hebbian adaptation. Its core process
is Spreading-Activation over Semantic Networks, which is a well-established theory
of semantic memory organization (Collins & Loftus, 1975; Cottrell & Small, 1983;
Quillian, 1966; Waltz & Pollack, 1985). According to this theory, each concept is
represented by a node in a semantic network, and semantically related nodes are
connected with weighted links. When a concept is processed, the appropriate node is
activated and the activation spreads along the network connections in a manner
proportional to the connection weights, gradually decreasing in strength over time. In
order to bring a word to awareness, its activation must exceed a threshold. Thus, the
spreading activation model provides a simple, transparent process that can be readily
matched with experimental data.
Before describing the details of SEMANT and its predictions, in the following
sections the empirical data on which it is based will be reviewed.
Silberman, Bentin and Miikkulainen 8
3. Human Performance Experiments
Two experiments were performed on human participants, in order to
characterize the interactions between episodic and semantic memory systems in some
detail. The first experiment showed that it is easier to form new episodic associations
between semantically related words compared with unrelated words. During an
incidental study phase, 10 randomly ordered Hebrew word pairs were repeated 20
times. In each trial, two words were displayed sequentially for 350 ms each and a
Stimulus Onset Asynchrony (SOA) of 700 ms. After an additional SOA of 800 ms
(relative to the onset of the second word), a single letter was presented and the
participant was requested to determine (pressing an alternative button) whether this
letter was or was not included in either of the preceding words (Kutas & Hillyard,
1988; Smith, Bentin & Spalek, 2001). Hence, episodic proximity was established by
having the participants store the two words together in working memory for 800
milliseconds. The shallow processing orientation task was selected in order to avoid
task-biased semantic activity at study and depth of processing effects for single words
(cf. Craik and colleagues, e.g., Craik & Lockheart, 1972; Craik & Tulving, 1975).
Following the study session in which the participants performed with an
accuracy of over 90%, the strength of the association between the words in each pair
was unexpectedly tested using a cued recall test and a free association test. In the cued
recall test, the first words of half of the studied pairs were presented as cues, and the
participants were requested to respond with the cue's associate. In the free association
test, the participants were presented with the first words of the other half of the pairs,
and the participants were asked to respond with a first free-associate. Correct cued
recall performance and responding with the episodically paired word in the free
Silberman, Bentin and Miikkulainen 9
association tests were considered evidence that an association between the two words
was incidentally formed during study. The effect of semantic relationship was tested
between subjects. One group of 32 participants was tested with semantically related
but not associated pairs (e.g. MILK-SOUP)1 and another equally sized group was tested
with semantically unrelated words (e.g. PAINT-FOREST). The performance of the 64
subjects is presented in Table 1.
Table 1: Percentage of cued recall and free association for pairs of semantically related and unrelated words.
Relatedness Cued Recall Free Associations
Related 38.8% 7.5%
Unrelated 19.4% 1.3%1
1 Based on 16 participants.
Both cued recall and free association revealed stronger associations between
semantically related pairs than for unrelated pairs [t(62)=2.8, p
Silberman, Bentin and Miikkulainen 10
This effect makes it easier to form associative connections during each new episode of
concomitant activation. Although the two accounts are not mutually exclusive, they
should affect cued recall performance in distinguishable ways. Pre-existent (weak)
associations should enhance associative strength at the very beginning of the episodic
associative learning process. Following this initial boost, however, pre-existent
association should not influence the learning curve. In contrast, facilitated associative
learning should be evident (at least up to a ceiling) at each recurrent learning episode.
Consequently, the effect of efficient spread of activation should manifest as a steeper
learning function for related than unrelated pairs.
In order to test the relative contribution of each of these two factors, in a
subsequent experiment the semantic relationship between the words in each pair
(within-subject factor) and the number of repetitions during the incidental learning
phase (between-subjects factor) were manipulated. Twenty-four pairs of exemplars
representing 24 semantic categories were selected. As in the previous experiment, the
words in each pair did not elicit one another in free association questionnaire in which
participants were requested to provide the first three associates in response to a cue
word. From the list of 24 related pairs, two study lists were prepared. Each list
comprised of 12 related pairs form the list above, and 12 unrelated pairs. The
unrelated pairs were scrambled pairings of the other 12 pairs. The related pairs in list
A were scrambled in list B and the unrelated pairs in list A were reorganized to form
the related set in list B. Four different groups of 24 participants were assigned to one
presentation (i.e., no repetition), 5, 10 or 20 presentations during an incidental
learning phase, during which they performed the letter search task. Following study,
the participants were unexpectedly presented with a cued recall test in which the first
Silberman, Bentin and Miikkulainen 11
word of all 24 pairs were presented as cues and the participants were requested to
respond with the paired associated word from the study phase.
All participants performed the letter search task at study with accuracy above
85%, suggesting that they paid attention to the words. Cued recall was better for
semantically related than unrelated words for all groups, and the semantic advantage
increased with increased repetition (Figure 1). Mixed-model ANOVA showed that
both main effects were significant [F(1,92)=204, p
Silberman, Bentin and Miikkulainen 12
The second experiment, therefore, extended the results of the first experiment,
distinguishing between semantic and episodic contributions to the associative process.
Together, the results describe in detail how semantic relatedness facilitates forming
incidental episodic associations between words. Next, a computational model for this
process is presented, and predictions for further experiments derived.
4. Computational Model - SEMANT
The computational simulations were based on SEMANT (Semantic and
Episodic Memory of Associations using Neural neTworks), a biologically-motivated
and psychologically real neural network model of the semantic system. Following an
overview of SEMANT, the semantic representations, the network architecture, and the
model dynamics will be described in this section.
4.1 Overview
SEMANT is a two-dimensional self-organizing map network where
semantically related concepts are represented by nodes that are closer to each other
than semantically unrelated nodes. Episodic associations are represented on this map
as direct connections among the nodes. Activation among nodes spreads along both
semantic and associative connections. Inspired by the notion of Hebbian learning, the
strength of an association between two conjointly activated nodes is enhanced when
the "wave" of activation spreading from one node overlaps with the activation
spreading from another node. Since spreading activation decays with distance from
the source, the closer the two concepts are on the semantic map, the greater the
overlap between their activations. Hence, when two semantically related words are
conjointly activated, the associative connection strength between them increases by a
large amount. On the other hand, because semantically unrelated words are further
Silberman, Bentin and Miikkulainen 13
apart on the semantic map, their concomitant activations overlap less and
consequently their associative connection strength increases by a smaller amount.
4.2 Input Representations
Word semantics was based on lexical co-occurrence analysis according to the
Hyperspace Analogue to Language (HAL) model of Burgess and Lund (1997). In this
model, high-dimensional numeric representations for words are formed based on the
context in which they occur in large text corpora. A moving window of several words
is placed around each word in the text. A co-occurrence matrix is calculated according
to how often each word occurs in the different positions in the window. The rows and
columns of this matrix are high-dimensional vector representations for each word. To
make them computationally tractable, the 100 dimensions with the highest variance
are used for the final representations. As has been shown in a number of studies, the
resulting HAL vectors capture the semantic of words fairly well: Semantically related
words have similar, closer, representations (Burgess & Lund, 1997; Li, 1999; 2000;
Li, Burgess & Lund, 2000; Lund, Burgess & Atchley, 1995; Lund, Burgess & Audet,
1996).
In the present study, HAL representations were 100 dimensions long and formed
based on the 3.8-million-word CHILDES database (Li, 1999; MacWhinney, 2000).
Forty-eight nouns were selected from this database to form 24 pairs of words. The
words in each pair were exemplars from the same semantic category. The words were
the English translation of the 48 Hebrew words used in the second of the human
experiments described above. In some cases, if a direct translation did not exist or the
translated word did not appear in the set of HAL representations, a similar English
word was chosen. Another 202 nouns were selected randomly from the same set of
Silberman, Bentin and Miikkulainen 14
representations in order to create a richer semantic neighborhood for the 48 words of
interest. The words are listed in the appendix.
4.3 Self-Organizing Maps
The SEMANT architecture is based on a Self-Organizing Semantic Map with
lateral connections (Miikkulainen, 1992; Miikkulainen, Bednar, Choe, and Sirosh,
2005; Ritter & Kohonen, 1989). The Self- Organizing Feature Map (SOM) is an
Artificial Neural Network (ANN) that classifies high-dimensional data and represents
their topological structure in a two-dimensional map (Kohonen, 1990). The algorithm
has two different versions of implementation. One version is biologically motivated
and, therefore, includes computations that are biologically plausible. The other
version is oriented towards practical applications, and the equations that govern it are
abstractions of the computations in the biological version. SEMANT is based on a
hybrid implementation that retains the computational efficiency of the abstract model
but includes psychologically verified mechanisms when they are crucial for the
performance of the model.
The SOM architecture is a feed-forward, two-layered neural network. The
input is presented on the first layer, while the second layer, consisting of a 2-D array
of nodes, self-organizes to represent the topology of the input space. In SEMANT, the
nodes of the output layer are interconnected with all-to-all lateral connections. Inputs
that are significantly different in their semantic features are mapped onto distinct
locations in the output space; similarly, inputs with many overlapping semantic
features are mapped onto nearby locations. The organization of the map, i.e. the
assignment of weight vectors to the nodes in the second layer, is formed in an
unsupervised, iterative process, driven by the statistics of the input's features.
Silberman, Bentin and Miikkulainen 15
The map's response to the input vector is determined by calculating the scalar
product of the weight vectors of each node and the input vector:
where sij is the response of node (i,j) in a two-dimensional map, x is the input vector
and wij,k is the weight between component k of the input vector on node (i,j). Each
iterative adaptation step consists of two stages: First, the node in the output layer that
responds most strongly to the current input vector is found. Second, the weight
vectors of the most responsive node and neighboring nodes are modified to become
more similar to the input vector. The neighborhood in which weight vectors will be
modified is defined around the most responsive node, with a radius that decreases
from nearly the size of the entire map down to zero as the self-organization
progresses. The weight vectors of nodes in the neighborhood are changed to become
closer to the input vector according to
where wij(t) is the weight vector of node (i,j) at time t, x(t) is the input vector at time
t, and ε(t) is the adaptation rate, which is usually decreased to zero with t. In this
process, the weight vectors of the map nodes become approximations for the input
vectors and the weight vectors of neighboring nodes become similar, which results in
an ordered map of the input space.
4.4 Semantic Maps
The SOM algorithm is commonly used to visualize high-dimensional inputs
with inherent metric properties (Kohonen 1990; 2000). Such inputs are typical to low-
,,∑=k
kkijij xws(1)
(2) ( ) ( ) ( ) ( ) ( )[ ] ,1 twtxttwtw ijijij −+=+ ε
Silberman, Bentin and Miikkulainen 16
level perception. However, when dealing with high-level processing such as
semantics in language and memory, the inputs are often discrete and not necessarily
metric. There are two major ways to represent semantics (Ritter and Kohonen, 1989).
According to the first, the representation is a feature vector, where bits represent a set
of fixed properties (big/small, has 2 legs/4 legs, etc.). According to the other, the
context of the word, e.g. the random representation of the word and its predecessor
and successor words, is used as the input. In both cases, Semantic Maps are created,
that is, meaningful maps of word semantics that express grammatical and semantic
relationships between words. Because the latter methodology can be automated, it can
be applied to large corpora such as the entire text of the Grimm fairy-tales (Honkela,
Pulkki and Kohonen, 1995). Therefore it was also adopted for SEMANT, using the
HAL approach for creating the representations.
Semantic maps have been successfully used in various investigations of the
semantic system addressing issues such as language acquisition, semantic priming,
semantic and episodic memory, and were used to document representation and
retrieval (Kohonen et al., 2000; Li, Farkas & MacWhinney, 2004; Lowe, 1997;
Mayberry, 2004; Miikkulainen, 1993; Scholtes, 1991). Because self-organizing maps
are based on Hebbian learning and maps in general are common in many parts of the
cortex (Knudsen, Lac & Esterly, 1987), self-organizing maps are most appealing as a
biologically plausible analogue of classic semantic networks (Spitzer, 1997).
4.5 SEMANT architecture
The core of the SEMANT model is a semantic map consisting of 250 nouns
organized in a 40 by 40 grid. The semantic map is extended to include all-to-all
unidirectional lateral connections. These connections represent potential associations
Silberman, Bentin and Miikkulainen 17
between two words. The strength of each connection is composed of semantic and
episodic components:
where Latij,uv is the connection weight from node (i,j) to node (u,v). The semantic
component represents the distance on the map, given by
where wij is the map's weight vector for node (i,j) and AMP, SLP, and SFT are free
parameters. 2The equation describes a reverse sigmoid with AMP defining its height,
SLP its slope and SFT its offset. Therefore, the strength of each lateral connection is
inversely proportional to the 100-dimensional distance between the nodes' weight
vectors. Initially, the episodic part of all lateral connections is zero. Hence, prior to
any learning of new associations, the lateral connections capture only the topographic
organization of the map, that is, the words' semantic relations.
The semantic map was organized in two phases. In the first phase, aimed
at producing a gross organization, the neighborhood size was decreased from 40 to 3
over 10,000 iterative adaptation step. In the second organization phase, aimed at fine
tuning the map's representations, the neighborhood size was decreased from 3 to 0
over additional 20,000 iterations. The adaptation rate (ε) was 0.5 throughout the
organization process but decreased linearly as a function of the Euclidian distance
square between the updated node and the most responsive node at each iteration.
2 There are a total of 19 free parameters in this model.
,,,, uvijuvijuvij EpiSemLat +=(3)
(4) ,
12,
⎟⎟⎠
⎞⎜⎜⎝
⎛+
=⎟⎠⎞⎜
⎝⎛ −− SLPSFTww
uvijuvije
AMPSem
Silberman, Bentin and Miikkulainen 18
4.6 SEMANT Dynamics
When a word is presented to SEMANT, an activity "bubble" develops
surrounding the node that represents the word in the network. The generated activity
wave spreads from the activated node according to synchronized recurrent dynamics.
At each time step, the input to each node is the sum of the activities of all nodes in the
previous time step, weighted by the lateral connections. The node's activity is limited
between the values 0 and 1 according to a sigmoid function
where
and ( )tSij is the activity of node (i,j) at time t.
When two words are presented to SEMANT during the same learning trial (τ),
the episodic weights adapt to encode a lateral connection (i.e., an association) between
them. Both activities spread independently over the map and the intersection of these
activations is summed over all the map's nodes and over all time steps. This sum is
added to the episodic component of the lateral connection between these words (only
in the direction that corresponds to the order of presentation) according to
where ( )tsPRMmn is the activation of node (m,n) at time t resulting from the presentation
of the prime word, which is represented in node (iPRM,jPRM), and ( )ts PRMmn is the
activation of node (m,n) at time t resulting from the presentation of the target word,
which is represented in node (iTGT,jTGT).
( ) ( ) ,1, ⎟⎠
⎞⎜⎝
⎛−= ∑
uvuvuvijij tsLatts σ(5)
( ) ( ) ,11
xex
−+=σ(6)
( ) ( ) ( ) ,)(),(min1,
,, ∑+=+tmn
TGTmn
PRMmnjijijiji tstsEpiEpi TGTTGTPRMPRMTGTTGTPRMPRM ττ(7)
Silberman, Bentin and Miikkulainen 19
When the distance between the words is small (indicating that the words are
strongly semantically related), the resulting activity waves overlap significantly and the
connection between them is considerably amplified at each co-occurrence. Thus, it is
easier for SEMANT to associate related words than unrelated words. Conceptually, this
method is an abstraction of Hebbian learning of episodic links, since the resulting
connection strength depends on the intersection of the activation waves of both words.
4.7 Computational Participants
In the simulation experiments, 12 different computational participants were
generated by self-organizing the semantic map each time from different random initial
starting points. The same sequence of input words was used for each participant (see
appendix for detailed simulation parameters).
The resulting maps learned to represent the semantic similarities in the data, but
differed in the details of how they were organized. For example, in the map of Figure 2,
food-related concepts are clustered in the bottom and body parts on top. The food and
body part clusters were prominent on other maps as well, but were shaped differently and
located in different regions of the map. In this sense, the 12 maps can be seen as different
individuals with roughly equivalent semantic systems. Simulated psychological
Figure 2: A self-organized semantic map. Semantically related words, such as those denoting body parts and foods, are mapped to adjacent nodes. Twelve different maps were organized from different random initial values, and were used to simulate different participants.
Silberman, Bentin and Miikkulainen 20
experiments were conducted in this group of computational participants in order to
explain the performance observed in human studies.
5. Validating SEMANT as a Model of Human Learning
The first simulation was aimed at verifying the psychological validity of
SEMANT, by replicating the human performance in the second experiment described
above. Again, this experiment demonstrated that it is easier to form associations
between semantically related than unrelated words, and that this advantage increases
with repetitive presentation.
5.1 Simulation 1
5.1.1 Experimental Setup
Out of the 24 semantically related pairs on the semantic map 12 pairs were
chosen with a Euclidean distance between HAL representations shorter than average
(0.88) but larger then a pre-determined threshold (0.35). This selection insured that
the words in each pair were semantically related, but not strongly enough to be always
recalled regardless of any episodic associations. In addition, the other 12 pairs were
randomly re-matched to form 12 pairs of semantically unrelated words. The
simulation procedure was replicated 12 times using a different map each time and
statistical analysis was performed on the data.
5.1.2 Procedure
In each trial during the simulated study phase, SEMANT was given two
words one after the other at an appropriate SOA (i.e., the first word was given at
time step 0 and the second word only at time step 3). The absolute time scale of
the network is arbitrary and was adjusted to fit the data. Each of the 24 pairs (12
related and 12 unrelated) was presented once. Because episodic factors do not
Silberman, Bentin and Miikkulainen 21
affect how activation spreads in the modeled semantic network during the learning
phase, the episodic associative strength resulting from multiple presentations was
calculated by multiplying the result of a single presentation by the number of
repetitions, which varied from 1 to 30.
In each trial during the test phase, the first word in each of the studied pairs
was presented again to SEMANT. The resulting activation wave spread based on the
same dynamics as during the study phase, except that both semantic and episodic
components of the lateral connections now affected propagation. The activity
continued to spread until one of the nodes representing a word reached a pre-
determined activity threshold (0.98). This word was then emitted as output. If no
word-node reached the threshold within eight time steps (at which time the activation
wave had usually decayed to a negligible level), the word with the highest activation
was emitted as an answer. The latter scenario simulated the situation in which the
participant could not recall any word and would answer with the word that first came
to mind.
5.1.3 Results and Discussion
As evident in Figure 3, SEMANT successfully replicated the pattern of results
found in human subjects. At early stages, the semantically related pairs were
associated faster than unrelated pairs. After about 10 repetitions, a ceiling effect
reduced this pace, such that their advantage over the unrelated pairs stops growing. In
contrast, the pace of learning unrelated pairs was relatively constant throughout the
repetitions.
It is important to emphasize that due to the sigmoid transfer function
(Equations 5 and 6) as well as the recurrent dynamics, SEMANT's response changes
nonlinearly with increasing repetitions. Moreover, the performance in the simulated
Silberman, Bentin and Miikkulainen 22
cued-recall test depends nonlinearly on the lateral connection strengths. Hence, the
nearly linear learning curves in SEMANT (until the ceiling effect) cannot stem merely
from the linear way in which multiple repetitions were modeled (Section 4.1.2).
Instead, they accurately represent an equal contribution of each learning trial to the
correct cued-recall performance.
Statistical analysis of the data revealed that both main effects, i.e. the effect of
semantic relatedness and the effect of number of repetitions, as well as the interaction
between them were reliable [F(1,11)=104.4 P
Silberman, Bentin and Miikkulainen 23
6. Implicit Asymmetry in Forming Associations
Unlike semantic relationships, associations between words are directional. In
free association questionnaires, for most pairs the participants would reply with word
B after A with a different probability than the other way around (Koriat, 1981). In
SEMANT, explicit asymmetry is achieved by the unidirectional lateral connections,
which represent the association between two words in the map. However, SEMANT
demonstrates an additional asymmetry that is implicit: It is sometimes easier to form
an association between two words in one direction than in the opposite direction even
before any episodic information is taken into account. Simulation 2 was aimed at
quantifying this phenomenon.
6.1 Simulation 2
6.1.1 Experimental Setup
The density of the semantic neighborhoods of the words used in Simulation 1
was determined by counting the number of words (out of the 250 total words in
SEMANT's lexicon) that were within a fixed 100-dimensional distance (0.4) in their
HAL representations. Based on this count 3 related and 3 unrelated pairs were
selected in which the difference in the semantic densities of the two words in a pair
was the greatest. These 6 pairs were used in this simulation. As before, the simulation
procedure was replicated 12 times using a different map each time and statistical
analysis was performed on the data.
6.1.2 Procedure
As in Simulation 1, the six pairs were presented one word at a time. First, the
pairs were presented in the forward direction, from the word with the sparse semantic
neighborhood to the word with the dense semantic neighborhood in each pair. Then,
Silberman, Bentin and Miikkulainen 24
the network was reset to its original state and the entire procedure was repeated with
the pairs presented in the opposite order (from dense to sparse). The "cued-recall"
performance of the two directions was compared statistically.
6.1.3 Results and Discussion
As shown in Figure 4, when the pairs were presented in the forward direction
(sparse neighborhood → dense neighborhood) associations were learned faster than
when the order of presentation was reversed [F(1,11)=24.8 p
Silberman, Bentin and Miikkulainen 25
to semantic memory than to establish a link between two formerly unconnected words
already existing in semantic memory. This result is consistent with the asymmetrical
directionality prediction of SEMANT if newly learned words are assumed to be
poorly embedded into their semantic neighborhood and, therefore, to have a sparser
semantic neighborhood than familiar words do.
7. Modeling Associative Learning in Schizophrenia
Based on Collins and Loftus’s (1975) theory of spreading activation, Spitzer
(1997) suggested a model of schizophrenic thought disorder (STD) where the
activation wave over the semantic network of STD patients spreads faster and farther
away from origin than that of normal participants. Such atypical activation may
explain experimental results in STD patients, who show stronger direct priming (e.g.
between the words COW and MILK) as well as more indirect, mediated semantic
priming (e.g. between the words BULL and MILK, mediated by the word COW) than
normal subjects. It may also explain the clinical phenomenon of loose, oblique and
derailed associations. There are at least two possible accounts for the aberrant
spreading of activation in the STD semantic system: (1) the total amount of activation
0%
20%
40%
60%
80%
100%
0 5 10 15 20
Repetitions
Cor
rect
Cue
d R
ecal
l
Backward-RelatedForward-RelatedBackward-UnrelatedForward-Unrelated
Figure 4: Average percentages of correct recall demonstrated by the model for related and unrelated word pairs in forward (sparse to dense) and backward (dense to sparse) directions. The model predicts that learning is easier from sparse to dense neighborhoods.
Silberman, Bentin and Miikkulainen 26
may be higher in STD patients than in typical population, and (2) the activation may
be shallower in STD patients but more widespread. In Simulations 3A and 3B these
two possible accounts are tested in SEMANT. By comparing the performance in each
of the simulations with those of human participants it is possible to obtain insight into
what might be causing semantic disorders in STD patients.
7.1 Simulation 3A
7.1.1 Experimental Setup and Procedure
This simulation tested the effect of increasing the overall amount of activation
on associative learning. The experimental setup and procedures of Simulation 1 were
replicated, except that the amplitude of the spreading activation wave was relatively
elevated. As described above (Equation 4), the distance between the word
representations (nodes) determines how strongly they are linked according to a
reverse sigmoid function (Equation 4). This sigmoid was elevated by multiplying the
SFT parameter of Equation 4 by a factor of two. As a result, the area underneath the
curve increased but the shape of the sigmoid remained similar to that used in
Simulation 1 (Figure 5). The psychological interpretation of this change is that the
stronger and unfocused activation wave results from a higher level of excitability in
the system. As in Simulation 1, the simulation procedure was replicated 12 times
using a different map each time and statistical analysis was performed on the data.
Silberman, Bentin and Miikkulainen 27
7.1.2 Results and Discussion
As shown in Figure 6, running the simulation, SEMANT predicted that the
associative recall performance of STD patients in both related and unrelated
conditions would improve, at least up to about 10 repetitions [F(1,11)=58.9 P
Silberman, Bentin and Miikkulainen 28
associations, in the sense that semantically unrelated words are easier for STD
patients to associate than for matched control participants.
0%
20%
40%
60%
80%
100%
0 5 10 15 20
Repetitions
Cor
rect
Rec
all
Normal-RelatedNormal-UnrelatedSTD-RelatedSTD-Unrelated
Nevertheless, the superior associative recall performance of STD patients
predicted by simulation 3A for both related and unrelated pairs is intriguing.
Notwithstanding the stronger semantic priming effect, STD patients do not usually
succeed in memory tests better than normal participants, especially with incidentally
learned associations. Note that the better cued-recall for STD patients resulting from
Simulation 3A is a result of higher activation levels in general. To address this
problem and explore the second account for the unusual priming effects observed in
STD patients, in Simulation 3B a shallower but broader sigmoid was used. Such a
sigmoid generates unfocused activation without increasing its overall amount. The
goal was to see whether more loose associations would result without a significant
increase in overall recall performance.
Figure 6: Averaged percentages of correct recall demonstrated by SEMANT for normal participants and for STD patients with elevated activity. Performance for STD patients is elevated for both related and unrelated pairs and the difference between the two conditions is diminished, suggesting that mediated associations in STD may result from enhanced semantic connections.
Silberman, Bentin and Miikkulainen 29
7.2 Simulation 3B
7.2.1 Experimental Setup and Procedure
The setup and procedures of this simulation were similar to that of Simulation
3A except that the sigmoid was normalized such that the sum of the semantic
components of the lateral connections was identical to that used in Simulation 1
(Figure 7). Thus, the total strength of the lateral connections in the simulated STD
was the same as in the normal case; keeping the overall activation at the same level in
both groups. However, the connection profile was much wider, representing an
unfocused association system. As in all our previous simulations, the simulation
procedure was replicated on 12 different maps and analyzed statistically.
7.2.2 Results and Discussion
Unfocused activation was implemented in SEMANT by making the sum of the
semantic components of the lateral connections the same in STD and in the typical
case. As can be seen in Figure 8, this simulation resulted in lower percentage of cued
recall in related pairs while the performance of unrelated pairs was higher than in the
Figure 7: Semantic relatedness as a function of the distance between word representations in normal participants (solid line) and in STD patients (dashed line), modeled by a shallower and wider curve. Such a curve corresponds to higher unfocused associations.
Silberman, Bentin and Miikkulainen 30
normal case. Statistical analysis (ANOVA) did not show a statistically reliable
difference between normal and STD conditions [F(1,11)=0.086 P>0.75], but resulted
in a significant relatedness effect [F(1,11)=193.893 P
Silberman, Bentin and Miikkulainen 31
condition. In contrast, assuming equal excitability in both groups, but spread more
diffusely in the STD participants (simulated by a sigmoid with lower amplitude but
further reaching influence, Figure 7) resulted in better associative learning of
unrelated pairs but reduced associative learning of related pairs. This result can be
useful in uncovering the source of thought disorders in schizophrenic patients. Indeed,
preliminary data from an ongoing study in human STD patients and normal
individuals suggests that the second model is better supported by real data (Bentin et
al., in progress).
8. General Discussion
The goal of the present study was to determine how semantic and episodic
factors interact in learning new incidental associations between words. Empirical
studies with human participants showed that semantically related words are associated
more easily than unrelated words, and that this difference is primarily due to higher
associative strength added by each episode of co-occurrence. A computational model,
SEMANT, was developed to account for this phenomenon and to suggest a more
general mechanism for word association. An initial simulation demonstrated that the
model is psychologically valid: The computational process of forming new
associations between related and unrelated words accurately replicated human
performance. Subsequent simulations generated predictions about associative learning
between different types of words and modeled different accounts for the peculiar
associative learning in schizophrenic patients with thought disorders. These model-
derived predictions (as well as others) can be tested empirically in future experiments
with human subjects. The model has also broader implications for the study of
memory, as will be discussed in this section.
Silberman, Bentin and Miikkulainen 32
SEMANT suggests that both semantic relationship and episodic associations
can be implemented in a single network structure using two types of representations.
Semantic relationship is expressed as proximity in a high-dimensional features' space
spanned by the numerical representations of the concepts. Episodic associations are
represented by arbitrary "physical" connections between the nodes that represent the
words. Both types of relations are implemented simultaneously in a self-organizing
semantic map with lateral connections. Based on such structure, SEMANT showed
that semantic relationship facilitates learning new associations. This facilitation
emerges in a natural, mechanistic manner, without involving top-down intentional
processes. It is achieved through a process similar to Hebbian learning, where
connections are strengthened based on intersections of spreading activation waves
over a semantic map.
In SEMANT, the episodic connections do not affect the organization of the
semantic map. This modeling approach reflects the assumption that different rules
govern the organization of the semantic and episodic memories: The semantic map is
organized according to outside (possibly sensory) information and is not affected by
the episodic connections. Indeed preliminary experiments show that newly formed
episodic associations do not affect the semantic memory, i.e. they do not bring entire
neighborhoods closer together (Silberman et al. 2005). Such effects may be possible
at very long time scales, which are currently outside the scope of the model.
The basic principles underlying SEMANT (map-like memory organization,
spreading activation, and Hebbian adaptation) are well established in the literature and
have been used to model several other phenomena. The fact that these principles can
also account for word associations suggests that they may, indeed, represent general
mechanisms of cognition. The predictive power in the model was demonstrated in a
Silberman, Bentin and Miikkulainen 33
number of cases. One prediction was that it is easier to form associations for sparse
to dense semantic neighborhoods than the other way around; another was that
excessive mediated associations in STD could be due to elevated or diffuse spreading
activation process, and these mechanisms can be distinguished in the data. A third
prediction is that a strong association between semantically unrelated words (e.g.
BEER - DAISY) facilitates forming new associations between other words from the
same semantic neighborhoods (e.g. WINE - TULIP). This is because each of the new (to
be associated) words activates its own semantic neighborhood, including the previously
associated words. The pre-existent episodic association mediates the spread of activation
between the new conjointly presented exemplars, thus enhancing the overlap of their
activations This prediction has already been validated empirically in human
(Silberman, et al., 2005).
The asymmetric nature of word associations is difficult to explain with
computational models that rely on distances between high-dimensional numeric
representations. One atypical example is Plaut's (1995) model, which can, in
principle, capture asymmetric associations between word pairs based on the relative
frequency of the two directions of presentation (although such behavior has not yet
been demonstrated in that model). Similarly, SEMANT is based on perfectly
symmetric foundations such as high-dimensional vectors and the self- organization
algorithm that establishes the semantic map. Nonetheless, SEMANT demonstrates
two kinds of asymmetries (as was discussed in section 6). The first asymmetry is
explicit; it is achieved by unidirectional lateral connections that are implemented on
top of the symmetric organization of the semantic map. These connections make it
possible to have asymmetric associations between two words, based on different
episodic experiences in the two directions. The second asymmetry is implicit
emerging from the non-uniform distribution of concepts in the high-dimensional
Silberman, Bentin and Miikkulainen 34
space. This distribution makes spreading activation asymmetric between two points
that would otherwise be equidistant in the semantic space. By doing so, SEMANT
suggests a mechanism based on which asymmetric behavioral phenomena can emerge
from seemingly symmetric building blocks.
SEMANT demonstrates implicit asymmetry because the weight vectors in the
semantic map are distributed non-homogeneously in the 100-dimensional space.
However, factors other than semantics may influence this distribution and
consequently influence the implicit asymmetry as well. As described in section 4, the
density of the weight vectors in the semantic map is determined during the self-
organization process and reflects how the numerical HAL representations of the 250
words are distributed. Lund et al. (1995, 1996) argued that HAL representations
reflect primarily the semantic features of the words. Therefore the semantic map in
SEMANT is likely to reflect primarily the semantic relationships between words prior
to the episodic learning. However, other factors such as word frequency and word
familiarity may be confounded in HAL representations as well, and therefore may
affect the semantic density and implicit asymmetry in SEMANT. Additional studies
are necessary to disentangle the putative contribution of such factors.
SEMANT also suggests that the nodes unassigned with words in the semantic
map are important, since they serve as a mediating substrate for spreading activation.
Any system organized in an unsupervised fashion to accommodate an unknown
number of items (such as the human semantic system) should be expected to have a
much larger capacity than is usually needed. In semantic maps, the number of nodes
in the map is much larger than the number of represented words (e.g. in SEMANT,
250 nodes are embedded in 402 nodes, resulting in 6.4 nodes per word on average).
However, due to the self-organization algorithm, these unassigned nodes are not
Silberman, Bentin and Miikkulainen 35
distributed uniformly in the high-dimensional space, but rather represent the densest
areas of the input space. Consequently, they help magnify the effects that are due to
the statistical properties of the input. An example of such an effect is the asymmetry
in the learning efficiency between two directions of word pairs, as demonstrated in
Simulation 2.
While the similarity between human performance and the computer
simulations suggest that SEMANT is a valid psychological model, it is still a high-
level abstraction of the underlying biological mechanisms. For example, recall that
during the simulation of the learning phase, two words are presented to SEMANT
with a certain SOA and the two activation waves spread independently. The lack of
interaction between the two spreading waves is a computational abstraction that could
be elaborated in a more biologically plausible model. More specifically, in order for
two (or more) waves to spread without interaction within the same system, the waves
can be labeled, or marked, to carry the information to identify their source throughout
their propagation. Such a mechanism is central in the marker-passing extensions to
spreading activation (Charniak, 1983; Hendler, 1988). One biologically plausible
marker passing mechanism is the synchronized spiking model (Choe & Miikkulainen,
1998; Miikkulainen et al., 2005; Shastri and Ajjanagadde, 1993; von der Malsburg,
1986). If two activations spike at different phases, the information regarding their
source is preserved. On the other hand, if the enhancement of the association strength
between the two words is done in real-time during the propagation of the waves, and
if the sensory input that created the activation is accessible throughout the
propagation, a marking scheme may not be needed: The nodes where the waves
originated can be identified because they respond maximally to the sensory input.
However, it may not be realistic to assume that two (or more) activation waves may
Silberman, Bentin and Miikkulainen 36
propagate through the same system (semantic memory) without interaction. Hence, a
more plausible assumption would be that waves that originate from different sources
interact in some manner and that the propagation of one affects the propagation of the
others. Changing SEMANT to support unlabeled interacting activation and still
replicating human behavior is an interesting future research line.
9. Conclusion
Based on detailed studies with human participants, a computational model of
forming incidental associations between words was developed. The model suggests
that this process could be based on spreading activation on a laterally connected self-
organizing map that combines episodic and semantic factors. It also demonstrates how
such associations could be asymmetric and how the associative process can be
impaired in STD patients. These results constitute a promising first step towards a
computational theory of associative thinking in human.
Silberman, Bentin and Miikkulainen 37
Acknowledgements
We are grateful to Ping Li and Curt Burgess for providing the HAL vectors. This research
was supported in part by the National Science Foundation grant IIS-981147 and National
Institute of Mental Health Human Brain Program grant R01 to Risto Miikkulainen, and by a
German-Israeli Science Foundation grant I-771-250.4/2002 to Shlomo Bentin.
Silberman, Bentin and Miikkulainen 38
References
Bower, G. H. (1961). Application of a model to paired-associate learning. Psychometrika, 26, 255-280
Burgess, C. & Lund, K. (1997). Modelling parsing constraints with high-dimensional context space. Language and Cognitive Processes, 12, 1-34.
Charniak, E. (1983). Passing markers: {a} theory of contextual influence in language comprehension. Cognitive Science, 7, 171-190.
Chiarello, C., Burgess, C., Richards, L., & Pollock, A. (1990). Semantic and associative priming in the cerebral hemispheres: Some words do, some words don’t… sometimes, some places. Brain and Language, 38, 75-104.
Choe, Y. & Miikkulainen, R. (1998). Self-organization and segmentation in a laterally connected orientation map of spiking neurons. Neurocomputing, 21, 139-157.
Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407-428.
Cottrell, G. W. & Small, S. L. (1983). A connectionist scheme for modeling word sense disambiguation. Cognition and Brain Theory, 6, 89-120.
Craik, F.I.M., & Lockheart, R.S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684.
Craik, F.I.M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104, 268-294.
Dagenbach, D., Horst, S., and Carr, T. H. (1990). Adding new information to semantic memory: How much learning is enough to produce automatic priming? Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 581-591.
Epstein, M. L., Phillips, W. D., & Johnson, S. J. (1975). Recall of related and unrelated word pairs as a function of processing level. Journal of Experimental Psychology: Human Learning and Memory, 104(2), 149-152.
Farah, M., & McClelland, J. (1991). A computational model of semantic memory impairment: Modality specificity and emergent category specificity. Journal of Experimental Psychology: General, 120, 339–357
Guttentag, R. (1995). Children's associative learning: automatic and deliberate encoding of meaningful associations. American Journal of Psychology, 108(1), 99-114.
Hendler, J. A. (1988). Integrating Marker Passing and Problem Solving: {A} Spreading Activation Approach to Improved Choice in Planning. Hillsdale, NJ: Erlbaum.
Hinton, G. E. (1990). Mapping part-whole hierarchies into connectionist networks. Artificial Intelligence, 46, 47-75.
Hinton, G., & Shallice, T. (1991). Lesioning an attractor network: Investigations of acquired dyslexia. Psychological Review, 98, 74–95.
Honkela, T., Pulkki, V., & Kohonen, T. (1995). Contextual relation of words in Grimm tales, analyzed by self-organizing map. Proceedings of international conference on artificial neural networks (pp. 3-7).
Knudsen, E. I., Lac, S., & Esterly, S. D. (1987). Computational maps in the brain. Annual Review of Neuroscience, 10, 41-65.
Kohonen, T. (1990). The self-organizing map. Proceedings of the IEEE (pp. 1464-1478). Kohonen, T., Kaski, S., Lagus, K., Salojrvi, J., Paatero, V., & Saarela A. (2000). Self-
organization of a massive document collection. IEEE Transactions on Neural Networks, 11, 574-585.
Kohonen, T., Kaski, S., Lagus, K., Salojrvi, J., Paatero, V., & Saarela A. (2000). Self-organization of a massive document collection. IEEE Transactions on Neural Networks, 11, 574-585.
Silberman, Bentin and Miikkulainen 39
Koriat, A. (1981). Semantic facilitation in lexical decision as a function of prime-target association. Memory and Cognition, 9, 587-598.
Kutas M and Hillyard SA (1988): An electrophysiological probe of incidental semantic association. Journal of Cognitive Neuroscience, 1: 38-49.
Li, P. (1999). Generalization, representation, and recovery in a self-organizing feature-map model of language acquisition. Proceedings of the 21st Annual Conference of the Cognitive Science Society (pp.308-313). Mahwah, NJ: Lawrence Erlbaum.
Li, P. (2000). The acquisition of tense-aspect morphology in a self-organizing feature map model. Proceedings of the 22nd Annual Conference of the Cognitive Science Society (pp.304-309). Mahwah, NJ: Lawrence Erlbaum.
Li, P., Burgess, C., & Lund, K. (2000). The acquisition of word meaning through global lexical co-occurrences. Proceedings of the 30th Child Language Research Forum (pp. 167-178). Stanford, CA: Center for the Study of Language and Information.
Li, P., Farkas, I., & MacWhinney, B. (2004) Early lexical development in a self-organizing neural network. Neural Networks 17, 1345-1362.
Lowe, W. (1997). Semantic representation and priming in a self-organizing lexicon. Proceedings of the 4th Neural Computation and Psychology Workshop (pp. 227-239). Springer-Verlag.
Lund, K., Burgess, C., & Atchley, R. A. (1995). Semantic and associative priming in high-dimensional semantic space. Proceedings of the 17th Annual Conference of the Cognitive Science Society (pp. 660-665). Mahwah, NJ: Lawrence Erlbaum.
Lund, K., Burgess, C., & Audet, C. (1996). Dissociating semantic and associative word relationships using high-dimensional semantic space. Proceedings of the 18th Annual Conference of the Cognitive Science Society (pp. 603-608). Mahwah, NJ: Lawrence Erlbaum.
MacWhinney, B. (2000). The CHILDES Project: Tools for Analyzing Talk. Volume 1: Transcription format and programs. Volume 2: The Database. Mahwah, NJ: Lawrence Erlbaum Associates.
Masson, M. E. J. (1995). A distributed memory model for semantic priming. Journal of Experimental Psychology: Learning, Memory & Cognition, 21, 3-23.
Mayberry, M. R. (2004). Incremental Nonmonotonic Parsing Through Semantic Self-Organization, Doctoral Dissertation, Department of Computer Sciences, The University of Texas at Austin. Technical Report AI-TR-04-310.
McGuire, W. J. (1961). A multiprocess model for paired-as-associate learning. Journal of Experimental Psychology, 62, 335-347.
McKoon, G., & Ratcliff, R. (1986). Automatic activation of episodic information in a semantic memory task. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 108-115.
McRae, K., & Boisvert, S. (1998). Automatic semantic similarity priming. Journal of Experimental Psychology: Learning, Memory, and Cognition , 24(3), 558-572.
Miikkulainen, R. (1992). Trace feature map: a model of episodic associative memory. Biological Cybernetics, 66, 273-282.
Miikkulainen, R. (1993). Subsymbolic Natural Language Processing: An Integrated Model of Scripts, Lexicon and Memory. Cambridge, MA: MIT press.
Miikkulainen, R., Bednar, J. A., Choe, Y., and Sirosh, J. (2005). Computational Maps in the Visual Cortex. New York: Springer.
Moss, H. E., Hare, M. L., Day, P., & Tyler, L. K. (1994). A distributed memory model of the associative boost in semantic priming. Connection Science, 6, 413-427.
Neely, J. H. (1991). Semantic priming effects in visual word recognition: A selective review of current findings and theories. In D. Besner, & G. W. Humphreys (Eds.), Basic Processes in Reading: Visual Word Recognition. Hillsdale, NJ: Lawrence Erlbaum.
Silberman, Bentin and Miikkulainen 40
Neely, J.H. & Durgunoglu, A. (1985). Dissociative episodic and semantic priming effects in episodic recognition and lexical decision tasks. Journal of Memory and Language, 24, 466-489.
Plaut, D. C. (1995). Semantic and associative priming in a distributed attractor network. Proceedings of the 17th Annual Conference of the Cognitive Science Society (pp. 37-42). Mahwah, NJ: Lawrence Erlbaum.
Plaut, D. C., & Shallice, T. (1993). Deep dyslexia: A case study of connectionist neuropsychology. Cognitive Neuropsychology, 10, 377–500.
Prior, A., & Bentin, S. (2003). Incidental formation of episodic associations: the importance of senential context. Memory & Cognition, 31(2), 306-316.
Prior, A., & Geffet, M. (2003). Mutual information and semantic similarity as predictors of word association strength: Modulation by association type and semantic relation. Proceedings of the European Cognitive Science Conference 2003 Mahwah, NJ: Lawrence Erlbaum Associates.
Quillian, M. R. (1966). Semantic Theory. Doctoral Dissertation, Carnegie Institute of Technology.
Ritter, H. & Kohonen, T. (1989). Self-organizing semantic maps. Biological Cybernetics, 61, 241-254.
Scholtes, J. C. (1991). Unsupervised learning and the information retrieval problem. Proceedings of the IEEE International Joint Conference on Neural Networks (pp. 95-100). Piscataway, NJ: IEEE Service center.
Schrijnemakers, J.M.C. & Raaijmakers, J.G.W. (1997). Adding new word associations to semantic memory: Evidence for two interactive learning components. Acta Psychologica, 96, 103-132.
Silberman, Y., Miikkulainen, R., & Bentin S. (2005). Associating unseen events: Semantically mediated formation of episodic associations. Psychological Science, 16, 161-166.
Seidenberg, M.S., & McClelland, J.L. (1989). A distributed developmental model of word recognition and naming. Psychological Review, 96, 523-568
Shastri, L. & Ajjanagadde, V. (1993). From simple associations to systematic reasoning: {a} connectionist representation of rules, variables and dynamic bindings using temporal synchrony. Behavioral and Brain Sciences, 16, 417-494.
Shelton, J.R., & Martin, R.C. (1992). How semantic is automatic semantic priming? Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 1191-1210.
Smith, M.C., Bentin, S. & Spalek, T.M. (2001). Is spreading of semantic activation truly automatic? The influence of task and response delay. Journal of Experimental Psychology: Learning Memory and Cognition, 27, 1289-1298.
Spitzer, M. (1997). A cognitive neuroscience view of schizophrenic thought disorder. Schizophrenia Bulletin, 23, 29-50.
Underwood, B. J., & Schultz, R. W. (1960). Response dominance and rate of learning paired associates. Journal of General Psychology, 62, 153-158.
von der Malsburg, C. (1986). A neural cocktail-party processor. Biological Cybernetics, 54, 29-40
Waltz, D. L. & Pollack, J. B. (1985). Massively parallel parsing: a strongly interactive model of natural language interpretation. Cognitive Science, 9, 51-74.
Wicklund, D. A., Palermo, D. S., & Jenkins, J. J. (1964). The effects of associative strength and response hierarchy on paired-associate learning. Journal of Verbal Learning and Verbal Behavior, 3, 413-420.
Silberman, Bentin and Miikkulainen 41
Appendix
Entire lexicon:
1. address
2. ale
3. answer
4. aquarium
5. back
6. bag
7. baker
8. balance
9. ball
10. band
11. band-aid
12. baseball
13. basement
14. bath
15. bathroom
16. bathtub
17. beard
18. bed
19. bee
20. bend
21. bin
22. birthday
23. blender
24. board
25. boom
26. bracelet
27. branch
28. bump
29. bush
30. butterfly
31. cabbage
32. camel
33. cards
34. carriage
35. carrot
36. castle
37. cheese
38. chin
39. chocolate
40. clay
41. coat
42. coffee
43. coke
44. collection
45. copy
46. cottage
47. cotton
48. counter
49. course
50. cow
51. cowboy
52. crack
53. crown
54. cube
55. dad
56. desert
57. dice
58. direction
59. dish
60. doctor
61. doll
62. dollar
63. door
64. downtown
65. dress
66. dresser
67. drill
68. drop
69. ear
70. earth
71. electricity
72. elevator
73. equipment
74. eye
75. factory
76. fall
77. farm
78. farmer
79. film
80. fireman
81. flag
82. flour
83. fly
84. fruit
85. game
86. garbage
87. gasp
Silberman, Bentin and Miikkulainen 42
88. ghost
89. gown
90. group
91. guess
92. guitar
93. gun
94. haircut
95. hammer
96. hamster
97. hand
98. helicopter
99. hippopotamus
100. horse
101. hut
102. ice
103. ice-cream
104. iris
105. ironing
106. jeans
107. jelly
108. job
109. joke
110. jungle
111. key
112. kiss
113. kit
114. knock
115. lamb-chop
116. lasagna
117. letter
118. lettuce
119. lie
120. lollipop
121. machine
122. mail
123. mane
124. marble
125. material
126. math
127. means
128. microphone
129. monster
130. mountain
131. nail
132. nap
133. neck
134. necklace
135. office
136. oil
137. pack
138. page
139. painting
140. pajamas
141. palm
142. panda
143. parade
144. parking
145. part
146. party
147. pause
148. pay
149. pear
150. pepper
151. piano
152. picnic
153. pill
154. pineapple
155. pipe
156. pitch
157. play-dough
158. police
159. popcorn
160. popsicle
161. potato
162. power
163. price
164. prince
165. prize
166. pudding
167. purse
168. raccoon
169. racing
170. railroad
171. rake
172. refrigerator
173. ringing
174. river
175. road
176. robin
177. rock
178. roof
179. rose
180. sack
181. saddle
182. sailor
183. salad
Silberman, Bentin and Miikkulainen 43
184. scarf
185. school
186. second
187. shape
188. sign
189. slap
190. sleep
191. slide
192. snow
193. soap
194. soda
195. son
196. song
197. soup
198. squash
199. squirrel
200. squirt
201. stairs
202. steam
203. step
204. stew
205. stone
206. store
207. strap
208. straw
209. strawberry
210. string
211. sum
212. sunglasses
213. sunshine
214. sweater
215. sweatshirt
216. syrup
217. table
218. tail
219. temperature
220. throat
221. tick
222. ticket
223. toe
224. towel
225. train
226. trash
227. tree
228. tricycle
229. trouble
230. truth
231. tuba
232. tuna
233. van
234. vanilla
235. violin
236. wall
237. war
238. wash
239. washcloth
240. waste
241. wedding
242. whale
243. wheat
244. whisper
245. wine
246. wire
247. wool
248. worm
249. yawn
250. zoo
Silberman, Bentin and Miikkulainen 44
Word pairs used in the simulations:
Semantically unrelated:
1. necklace - vanilla
2. painting - train
3. van - store
4. guitar - cabbage
5. pear - electricity
6. pepper - song
7. coat - violin
8. office - dress
9. camel - roof
10. oil - cow
11. wall - strawberry
12. carrot - bracelet
Semantically related:
1. bag - purse
2. eye - chin
3. butterfly - bee
4. gun - stone
5. lettuce - chocolate
6. snow - sunshine
7. doctor - baker
8. dice - cards
9. coffee - soda
10. bandage - pill
11. ball - doll
12. desert - river
4.1 Overview4.2 Input Representations4.3 Self-Organizing Maps4.4 Semantic Maps4.5 SEMANT architecture4.6 SEMANT Dynamics5. Validating SEMANT as a Model of Human Learning5.1 Simulation 15.1.1 Experimental Setup5.1.2 ProcedureIn each trial during the simulated study phase, SEMANT was given two words one after the other at an appropriate SOA (i.e., the first word was given at time step 0 and the second word only at time step 3). The absolute time scale of the network is arbitrary and was adjusted to fit the data. Each of the 24 pairs (12 related and 12 unrelated) was presented once. Because episodic factors do not affect how activation spreads in the modeled semantic network during the learning phase, the episodic associative strength resulting from multiple presentations was calculated by multiplying the result of a single presentation by the number of repetitions, which varied from 1 to 30.
6. Implicit Asymmetry in Forming Associations6.1 Simulation 26.1.1 Experimental Setup6.1.2 Procedure6.1.3 Results and Discussion7.1 Simulation 3A7.1.1 Experimental Setup and Procedure7.2 Simulation 3B7.2.1 Experimental Setup and Procedure
8. General Discussion Appendix